fix: Add local scripts (#6)

Co-authored-by: Nick <[email protected]>

.scripts/get_manifest_digest.sh

@@ -0,0 +1,11 @@
+#!/usr/bin/env bash
+
+# Get a manifest digest. Example inputs:
+# - docker.stackable.tech/stackable/hello-world:0.0.1-SNAPSHOT-stackable0.0.0-dev
+# - docker.stackable.tech/stackable/hello-world:0.0.1-SNAPSHOT-stackable0.0.0-dev-amd64
+set -euo pipefail
+
+# Note: `docker manifest push` currently outputs the same hash, but `manifest`
+# is experimental and the STDOUT is more likely to change than the structured
+# output.
+docker buildx imagetools inspect --format '{{println .Manifest.Digest}}' "$1"

README.md

@@ -17,263 +17,11 @@ particular step in a workflow.
 | Image Repo Digest                  | `docker.stackable.tech/stackable/kafka@sha256:917f800259ef4915f976...` |
 | Digest                             | `sha256:917f800259ef4915f976e93987b752fd64debf347568610d7f685d2022...` |
 
-## `build-container-image`
-
-> Manifest: [build-container-image/action.yml][build-container-image]
-
-This action builds a *single* container image using `docker buildx build`. It does the following work:
-
-1. Free disk space to avoid running out of disk space during larger builds.
-2. Build the image using `docker buildx build`, outputting the architecture specific tag.
-
-This action is considered to be the **single** source of truth regarding the image manifest tag.
-All subsequent tasks must use this value to ensure consistency.
-
-### Inputs and Outputs
-
-> [!TIP]
-> For descriptions of the inputs and outputs, see the complete [build-container-image] action.
-
-#### Inputs
-
-- `image-name` (eg: `kafka`)
-- `image-index-manifest-tag` (eg: `3.4.1-stackable0.0.0-dev`)
-- `container-file` (defaults to `Dockerfile`)
-- `build-context` (defaults to `.`)
-<!--
-TODO (@NickLarsenNZ): Allow optional buildx cache
-- `build-cache-username`
-- `build-cache-password`
--->
-
-#### Outputs
-
-- `image-repository-uri` (eg: `localhost/kafka`)
-- `image-manifest-tag` (eg: `3.4.1-stackable0.0.0-dev-amd64`)
-- `image-manifest-uri` (eg: `localhost/kafka:3.4.1-stackable0.0.0-dev-amd64`)
-
-[build-container-image]: ./build-container-image/action.yml
-
-## `build-product-image`
-
-> Manifest: [build-product-image/action.yml][build-product-image]
-
-<!-- markdownlint-disable-next-line MD028 -->
-> [!NOTE]
-> The build step is not concerned with registries, ports, paths to repositories, but still requires
-> a name. If the name does not contain a registry, `hub.docker.com` (?) is implied. Therefore,
-> `localhost` will be used as the registry so as to avoid accidental interactions with an unintended
-> registry.
->
-> Ideally, bake should be refactored to use `localhost` as the registry for the previously mentioned
-> reason (whether or not that is behind some option).
-
-This action builds a *single* container image using `bake`. It does the following work:
-
-1. Free disk space to avoid running out of disk space during larger builds.
-2. Build the image using `bake` which internally uses `docker buildx`.
-3. Temporarily retag the image to use `localhost` instead of `docker.stackable.tech/stackable`.
-4. Produce output values to be used in later steps.
-
-This action is considered to be the **single** source of truth regarding image index tag and image
-manifest tag. All subsequent tasks must use these values to ensure consistency.
-
-Currently, bake provides the following ouput in the `bake-target-tags` file:
-
-```plain
-docker.stackable.tech/stackable/kafka:3.4.1-stackable0.0.0-dev-amd64
-```
-
-Until bake supports the ability to specify the registry, this action will retag the image as:
-
-```plain
-localhost/kafka:3.4.1-stackable0.0.0-dev-amd64
-```
-
-### Inputs and Outputs
-
-> [!TIP]
-> For descriptions of the inputs and outputs, see the complete [build-product-image] action.
-
-#### Inputs
-
-- `product-name`
-- `product-version`
-- `image-tools-version`
-- `build-cache-username`
-- `build-cache-password`
-
-#### Outputs
-
-- `image-manifest-tag`
-
-[build-product-image]: ./build-product-image/action.yml
-
-## `publish-image`
-
-> Manifest: [publish-image/action.yml][publish-image]
-
-This action signs and publishes a *single* container image to the given registry. It does the
-following work:
-
-1. Tag the `source-image-uri` with the specified `image-registry-uti`, `image-repository`, and
-   `image-repository`.
-2. Push the container image to the specified registry.
-3. Sign the container image (which pushes the signature to the specified registry).
-4. Generate an SBOM via a syft scan.
-5. Attest an image with the SBOM as a predicate (which pushes the attestation to the specified
-   registry).
-
-### Inputs and Outputs
-
-> [!TIP]
-> For descriptions of the inputs and outputs, see the complete [publish-image] action.
-
-<!-- markdownlint-disable-next-line MD028 -->
-> [!IMPORTANT]
-> For multi-arch images, the `image-manifest-tag` should have the `-$ARCH` suffix, as the tag
-> without it should be reserved for the image index manifest which will refer to container images
-> for each architecture we will push images for.
-
-#### Inputs
-
-- `image-registry-uri`
-- `image-registry-username`
-- `image-registry-password`
-- `image-repository`
-- `image-manifest-tag`
-- `source-image-uri`
-
-#### Outputs
-
-None
-
-[publish-image]: ./publish-image/action.yml
-
-## `publish-index-manifest`
-
-> Manifest: [publish-index-manifest/action.yml][publish-index-manifest]
-
-This action creates an image index manifest, publishes it, and signs it. It does the following work:
-
-1. Create an image index manifest and link to each architecture in `image-architectures`.
-2. Push the image index manifest.
-3. Sign the image index manifest (which pushes the signature to the specified registry).
-
-### Inputs and Outputs
-
-> [!TIP]
-> For descriptions of the inputs and outputs, see the complete [publish-index-manifest] action.
-
-#### Inputs
-
-- `image-registry-uri`
-- `image-registry-username`
-- `image-registry-password`
-- `image-repository`
-- `image-index-manifest-tag`
-- `image-architectures`
-
-#### Outputs
-
-None
-
-[publish-index-manifest]: ./publish-index-manifest/action.yml
-
-## `run-pre-commit`
-
-> Manifest: [run-pre-commit/action.yml][run-pre-commit]
-
-This action runs pre-commit by setting up Python and optionally the Rust toolchain and Hadolint in
-the requested version. It requires a checkout with depth 0. It does the following work:
-
-1. Installs Python. The version can be configured via the `python-version` input.
-2. Optionally sets up the Rust toolchain and Hadolint.
-3. Runs pre-commit on changed files.
-
-Example usage (workflow):
-
-```yaml
----
-name: pre-commit
-
-on:
-  pull_request:
-
-jobs:
-  pre-commit:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout
-        with:
-          fetch-depth: 0
-          submodules: recursive
-      - uses: stackabletech/actions/run-pre-commit
-```
-
-### Inputs and Outputs
-
-> [!TIP]
-> For descriptions of the inputs and outputs, see the complete [run-pre-commit] action.
-
-#### Inputs
-
-- `python-version`
-- `rust`
-- `rust-components`
-- `hadolint`
-
-#### Outputs
-
-None
-
-[run-pre-commit]: ./run-pre-commit/action.yml
-
-## `shard`
-
-> Manifest: [shard/action.yml][shard]
-
-This action produces a list of versions for a product. This is to be used as a matrix dimension to
-parallelize builds. It does the following work:
-
-1. Reads the `conf.py`, filtering versions for the product
-2. Write the JSON array of version to `$GITHUB_OUTPUT` for use in a matrix.
-
-Example usage:
-
-```yaml
-jobs:
-  generate_matrix:
-    name: Generate Version List
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout
-      - id: shard
-        uses: stackabletech/actions/shard
-        with:
-          product-name: ${{ env.PRODUCT_NAME }}
-    outputs:
-      versions: ${{ steps.shard.outputs.versions }}
-
-  actual_matrix:
-    needs: [generate_matrix]
-    strategy:
-      matrix:
-        versions: ${{ fromJson(needs.generate_matrix.outputs.versions) }}
-    # ...
-```
-
-### Inputs and Outputs
-
-> [!TIP]
-> For descriptions of the inputs and outputs, see the complete [shard] action.
-
-#### Inputs
-
-- `product-name`
-
-#### Outputs
-
-- `versions`
-
-[shard]: ./publish-index-manifest/action.yml
+## Available Actions
+
+- [build-container-image](./build-container-image/README.md)
+- [build-product-image](./build-product-image/README.md)
+- [publish-image](./publish-image/README.md)
+- [publish-index-manifest](./publish-index-manifest/README.md)
+- [run-pre-commit](./run-pre-commit/README.md)
+- [shard](./shard/README.md)

build-container-image/README.md

@@ -0,0 +1,36 @@
+# `build-container-image`
+
+> Manifest: [build-container-image/action.yml][build-container-image]
+
+This action builds a *single* container image using `docker buildx build`. It does the following work:
+
+1. Free disk space to avoid running out of disk space during larger builds.
+2. Build the image using `docker buildx build`, outputting the architecture specific tag.
+
+This action is considered to be the **single** source of truth regarding the image manifest tag.
+All subsequent tasks must use this value to ensure consistency.
+
+## Inputs and Outputs
+
+> [!TIP]
+> For descriptions of the inputs and outputs, see the complete [build-container-image] action.
+
+### Inputs
+
+- `image-name` (eg: `kafka`)
+- `image-index-manifest-tag` (eg: `3.4.1-stackable0.0.0-dev`)
+- `container-file` (defaults to `Dockerfile`)
+- `build-context` (defaults to `.`)
+<!--
+TODO (@NickLarsenNZ): Allow optional buildx cache
+- `build-cache-username`
+- `build-cache-password`
+-->
+
+### Outputs
+
+- `image-repository-uri` (eg: `localhost/kafka`)
+- `image-manifest-tag` (eg: `3.4.1-stackable0.0.0-dev-amd64`)
+- `image-manifest-uri` (eg: `localhost/kafka:3.4.1-stackable0.0.0-dev-amd64`)
+
+[build-container-image]: ./action.yml

build-container-image/action.yml

@@ -85,13 +85,13 @@ runs:
         echo "IMAGE_ARCH=${IMAGE_ARCH}" | tee -a "$GITHUB_ENV"
 
         IMAGE_MANIFEST_TAG="${IMAGE_INDEX_MANIFEST_TAG}-${IMAGE_ARCH}"
-        echo "IMAGE_MANIFEST_TAG=${IMAGE_MANIFEST_TAG}" | tee -a $GITHUB_OUTPUT
+        echo "IMAGE_MANIFEST_TAG=${IMAGE_MANIFEST_TAG}" | tee -a "$GITHUB_OUTPUT"
 
         IMAGE_REPOSITORY_URI="localhost/${IMAGE_NAME}"
-        echo "IMAGE_REPOSITORY_URI=${IMAGE_REPOSITORY_URI}" | tee -a $GITHUB_OUTPUT
+        echo "IMAGE_REPOSITORY_URI=${IMAGE_REPOSITORY_URI}" | tee -a "$GITHUB_OUTPUT"
 
         IMAGE_MANIFEST_URI="${IMAGE_REPOSITORY_URI}:${IMAGE_MANIFEST_TAG}"
-        echo "IMAGE_MANIFEST_URI=${IMAGE_MANIFEST_URI}" | tee -a $GITHUB_OUTPUT
+        echo "IMAGE_MANIFEST_URI=${IMAGE_MANIFEST_URI}" | tee -a "$GITHUB_OUTPUT"
 
         echo "::group::docker buildx build"
         # TODO (@NickLarsenNZ): Allow optional buildx cache

build-product-image/README.md

@@ -0,0 +1,54 @@
+# `build-product-image`
+
+> Manifest: [build-product-image/action.yml][build-product-image]
+
+<!-- markdownlint-disable-next-line MD028 -->
+> [!NOTE]
+> The build step is not concerned with registries, ports, paths to repositories, but still requires
+> a name. If the name does not contain a registry, `hub.docker.com` (?) is implied. Therefore,
+> `localhost` will be used as the registry so as to avoid accidental interactions with an unintended
+> registry.
+>
+> Ideally, bake should be refactored to use `localhost` as the registry for the previously mentioned
+> reason (whether or not that is behind some option).
+
+This action builds a *single* container image using `bake`. It does the following work:
+
+1. Free disk space to avoid running out of disk space during larger builds.
+2. Build the image using `bake` which internally uses `docker buildx`.
+3. Temporarily retag the image to use `localhost` instead of `docker.stackable.tech/stackable`.
+4. Produce output values to be used in later steps.
+
+This action is considered to be the **single** source of truth regarding image index tag and image
+manifest tag. All subsequent tasks must use these values to ensure consistency.
+
+Currently, bake provides the following ouput in the `bake-target-tags` file:
+
+```plain
+docker.stackable.tech/stackable/kafka:3.4.1-stackable0.0.0-dev-amd64
+```
+
+Until bake supports the ability to specify the registry, this action will retag the image as:
+
+```plain
+localhost/kafka:3.4.1-stackable0.0.0-dev-amd64
+```
+
+## Inputs and Outputs
+
+> [!TIP]
+> For descriptions of the inputs and outputs, see the complete [build-product-image] action.
+
+### Inputs
+
+- `product-name` (eg: `kafka`)
+- `product-version` (eg: `3.4.1`)
+- `image-tools-version` (eg: `0.0.13`)
+- `build-cache-username` (required) <!-- TODO: make the cache optional -->
+- `build-cache-password` (required) <!-- TODO: make the cache optional -->
+
+### Outputs
+
+- `image-manifest-tag` (eg: `3.4.1-stackable0.0.0-dev-amd64`)
+
+[build-product-image]: ./action.yml

build-product-image/action.yml

@@ -113,4 +113,4 @@ runs:
 
         # Add the contents of the env variables to the GitHub output, so that it
         # can be used as action outputs
-        echo "IMAGE_MANIFEST_TAG=$IMAGE_MANIFEST_TAG" >> $GITHUB_OUTPUT
+        echo "IMAGE_MANIFEST_TAG=$IMAGE_MANIFEST_TAG" | tee -a "$GITHUB_OUTPUT"